Scalable epidemic message passing interface fault tolerance

Resilience and fault tolerance are challenging tasks in the field of high performance computing (HPC) extreme scale systems. Components fail more often such systems, results application abort. Adopting fault–tolerance techniques can be consistently detect failures continue application’s execution even if exist. A prominent parallel programming specification, message passing interface (MPI), as it would used to implement failure detection consensus algorithm this paper. Although MPI does not facilitate tolerant behavior, work presents a tolerant, matrix based algorithm. The proposed uses Gossiping. To failures, randomised pinging will applied during by using piggybacked gossip messages. In order achieve on system, failed processes’ information sent same messages all alive processes. was implemented framework is completely tolerant. exhibit process were detected global has achieved system.